GOVERNMENT RIGHTS The United States Government has certain rights in this invention pursuant to Contract No. DE-AC02-06CH11357 between the United States Government and UChicago Argonne, LLC, representing Argonne National Laboratory.
FIELD The present invention relates to tunable resistance coatings. More particularly, the invention relates to compositions of matter and methods of manufacture of W:AlOx and Mo:AlOx thin film having tunable resistance.
BACKGROUND This section is intended to provide a background or context to the invention that is, inter alia, recited in the claims. The description herein may include concepts that could be pursued, but are not necessarily ones that have been previously conceived or pursued. Therefore, unless otherwise indicated herein, what is described in this section is not prior art to the description and claims in this application and is not admitted to be prior art by inclusion in this section. Thin film materials of metal-metal oxides of nanocomposites can have many applications, including resistive layers for electronic applications, such as, for example, electron multipliers like microchannel plates, resistive memories, electro-chromic devices, biomedical devices and charge dissipating coatings on micro-electromechanical systems. A related application, U.S. Ser. No. 13/011,645, which is incorporated by reference herein, describes microchannel plate fabrication by atomic layer deposition (“ALD” hereinafter), which provide an example of how one can benefit from the tunable resistance coatings and methods of preparation described herein.
SUMMARY In one embodiment, thin layers of nanocomposite tungsten-aluminum oxide (W:Al2O3) can be prepared for various commercial purposes. In another embodiment molybdenum-aluminum oxide (Mo:Al2O3) with tunable resistivity can be prepared for various commercial purposes. These thin layers preferably were deposited using ALD by combining alternating exposures to disilane and/or tungsten and molybdenum hexafluoride for W deposited by ALD with alternating exposures of trimethyl aluminum and water for Al2O3 deposition by ALD. The film thicknesses were measured by using ex-situ ellipsometry, cross-sectional scanning electron microscopy, and transmission electron microscopy. The crystallinity and topography were examined using X-ray diffraction and atomic force microscopy, respectively. The composition of the composite layers was investigated by X-ray photoelectron spectroscopy, and the electrical conductivity was evaluated using current-voltage measurements. The thickness, composition, microstructure, and electrical properties of W:Al2O3 and Mo:Al2O3 nanocomposite layers were smooth and showed an amorphous nature via X-ray diffraction scans. The growth rate varied between 1.1-5 Å/cycle depending on the precursor ratio as well as sequence ALD precursors. The nature of ALD precursor dose pulses sequencing demonstrated a significant influence on the electrical properties of the desired W:Al2O3 or Mo:Al2O3 composite materials. To understand the film growth and electrical properties of W:Al2O3 and Mo:Al2O3 layers, in-situ quartz crystal microbalance measurements were performed during the W (Mo),Al2O3 and W (Mo):Al2O3 ALD. The electrical properties for various compositions of W (Mo):Al2O3 nanocomposite layer were studied and advantageous attributes identified for a variety of commercial applications.
BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 shows a schematic of resistivity ranges which can be achieved for materials by selected mixing of metal and insulating components by deposition methods;
FIG. 2(a) shows QCM measurements performed during Al2O3 ALD various precursors exposure steps; FIG. 2(b) shows calculated mass add-on QCM for the several Al2O3 ALD cycles; FIG. 2(c) shows calculated mass/ALD cycle of Al2O3; and FIG. 2(d) shows thickness of Al2O3 on Si(100) vs. ALD cycles;
FIG. 3(a) shows mass add-on on QCM during ALD W deposition in various precursor exposure steps; FIG. 3(b) shows calculated mass add on QCM for the several ALD W cycles; FIG. 3(c) shows calculated mass per ALD cycle of W; and FIG. 3(d) shows thickness of W on (10 nm) Al2O3 passivated Si(100) vs. ALD cycles;
FIG. 4 shows an X-ray diffraction pattern of a 600 Å layer of ALD Al2O3 (bottom scan) and 500 Å layer of ALD W (top scan);
FIG. 5(a) shows QCM during Al2O3—W—Al2O3 ALD step with mass uptake during Al2O3—W—Al2O3 precursor's exposure; FIG. 5(b) shows calculated mass from FIG. 5(a) per ALD cycles; FIG. 5(c) shows zoom-in of Al2O3—W-section for the first W cycle of mass onto Al2O3; and FIG. 5(d) shows zoom-in of —W—Al2O3 sections for first Al2O3 cycles on a W layer;
FIG. 6 shows mass add-on on QCM for W:Al2O3 composites grown with different % W ALD cycles;
FIG. 7(a) shows mass/ALD cycle and FIG. 7(b) shows actual W mass into the W:Al2O3 system vs. % W ALD cycles;
FIG. 8(a) shows growth rate of W:Al2O3 deposited on Si(100) vs. W % ALD cycles and FIG. 8(b) shows thickness of (30% W:70% Al2O3) ALD cycle composition case and all these layers were deposited with an ALD precursor sequence of [x(TMA-H2O)-y(Si2H6—WF6)];
FIG. 9(a) shows an X-ray diffraction scan of the W:Al2O3 system vs. W % ALD cycles layers which were deposited on fused quartz substrates and FIG. 9(b) shows an X-ray scan of (30% W:70% Al2O3) ALD cycle condition sample that was annealed at 400° C. in Ar;
FIG. 10(a) shows TEM analysis of (30% W:70% Al2O3) ALD cycle composite layer cap with 6 nm Al2O3:Cross section HRTEM micrograph; FIG. 10(b) shows a nano beam diffraction (NBD) image from a composite layer film area; FIG. 10(c) shows a High Angle Annular Dark Field (HAADF) STEM Tomography image of composite layer; and FIG. 10(d) shows a high resolution and high magnification cross-section TEM image;
FIG. 11 shows a sputter XPS depth profile of W:Al2O3 composite layer cap with 6 nm Al2O3; this layer was deposited with (30% W:70% Al2O3) ALD cycle composition grown at 200 C;
FIG. 12(a) shows sputter-XPS scans of (30% W:70% Al2O3) ALD cycles composite layer: W4f; FIG. 12(b) shows Al 2p; FIG. 12(c) shows O 1s; FIG. 12(d) shows C 1s; FIG. 12(e) shows F 1s; and FIG. 12(f) shows Si 2p;
FIG. 13(a) shows an XPS scans of an as grown W:Al2O3 sample with of (30% W:70% Al2O3) ALD cycles and FIG. 13(b) after 400° C. annealing in 300 sccm Ar for 4 hours at pressure of 1 Torr;
FIG. 14(a) shows microstructure and conformality of the W:Al2O3 system: SEM W:Al2O3 on micro-capillary glass plate substrates; FIG. 14(b) shows a high magnification SEM image from the cross section of glass-capillary pores which shows the presence of a thin layer in the area where micro-capillary glass plate was uneven; FIG. 14(c) shows a cross section of one of the micro capillary pore walls which show W:Al2O3 deposited uniformly; and FIG. 14(d) shows an AFM scan on the collated inner surface of the one of the pores;
FIG. 15(a) shows mass uptakes on QCM during W:Al2O3 composite layer grown with (25% ALD W:75% ALD Al2O3)=3x(H2O-TMA)-1x(Si2H6—WF6) precursor exposures sequence; ALD precursors dose vs. mass uptake; FIG. 15(b) shows calculated mass uptake after each completion of cycle vs. ALD cycle and; FIG. 15(c) shows one super cycle of 3x(H2O-TMA)-1x(Si2H6—WF6) vs. dose time; and FIG. 15(d) shows calculated mass per ALD cycle vs. ALD cycle;
FIG. 16 shows mass add on QCM during W:Al2O3 composite layer grown with (25% W:75% Al2O3) ALD cycle condition with variation in the precursor dose sequence in ALD reactor;
FIG. 17(a) shows QCM mass add on for one super cycle (25% W:75% Al2O3) ALD cycle condition for various ALD precursor sequence: THSW; FIG. 17(b) shows THWS; FIG. 17(c) shows HTSW; and FIG. 17(d) shows HTWS;
FIG. 18 shows resistivity of W:Al2O3 composite layers as a function of % W ALD cycles with all the other parameters kept identical and the sequence of precursor=[x(TMA-H2O):y(Si2H6—WF6)] where x and y varied between 0-100;
FIG. 19 shows transverse (⊥) and longitudinal (∥) resistivity data for the as deposited and 400° C. annelid Ar samples of W:Al2O3 composite layer deposited with (30% ALD W:70% ALD Al2O3). =x(TMA-H2O):y(Si2H6—WF6);
FIG. 20(a) shows an IV curve of a film; FIG. 20(b) shows log(J/E) vs. log (E); FIG. 20(c) shows on a semi-log scale J/E vs. E1/2; and FIG. 20(d) shows data fitting to FIG. 20(c) at low field (top graph FIG. 20(d)(1) with linear fit), and high field (bottom graph FIG. 20(d)(2) with exponential fit);
FIG. 21(a) shows ln(R) vs. Temperature and FIG. 21(b) shows log(resistivity) vs. temperature for W:Al2O3 composite layer deposited with 30% W ALD cycle condition;
FIG. 22 shows an X-ray diffraction scan of an as-deposited MoAlOx layer having 8% Mo ALD cycles in an Al2O3 layer with amorphous structure;
FIG. 23(a) shows AFM images of an ALD MoAlOx tunable resistance coating prepared using 8% Mo ALD cycles in a series of Al2O3 ALD cycles; and FIG. 23(b) shows an image of a surface view of the coating;
FIG. 24(a) shows a top down XTEM image of an ALD MoAlOx tunable resistance coating prepared using an 8% Mo cycle with a thickness of 760 Angstrom on a Si substrate; and FIG. 24(b) shows a corresponding bright field TEM image;
FIG. 25(a) shows a cross section XTEM image of an ALD MoAlOx tunable resistance coating prepared using 8% Mo cycles and 760 Angstrom thickness on an Si substrate; FIGS. 25(b) and 25(c) show higher magnification views of FIG. 25(a);
FIG. 26(a) shows a photograph of MoAlOx films on glass and prepared with 8% Mo with a thickness of 760 Angstrom; FIG. 26(b) shows the film on a 300 mm Si wafer;
FIG. 27(a) shows a thickness contour map for 400 cycles with 5% Mo; FIG. 27(b) shows refractive index for FIG. 27(a); FIG. 27(c) shows contours for 8% Mo; FIG. 27(d) shows refractive index for 27(c); FIG. 27(e) contours for 12% Mo and FIG. 27(f) refractive index for 27(e);
FIG. 28 is a schematic diagram of location of 26 kapton discs (back face electrodes) affixed to a Mo-coated 300 nm Si wafer prior to ALD MoAlOx film deposition with 8% Mo couple is Al2O3;
FIG. 29(a) shows a current/voltage plot of a MoAlOx film prepared using 8% Mo cycles with measurements performed at 26 locations on a 300 nm wafer (see FIG. 28); FIG. 29(b) shows a magnified portion of FIG. 29(a);
FIG. 30 shows transverse resistivity of a MoAlOx film prepared using 8% Mo cycles with resistivity obtained from I/V measurements at 26 locations on a 300 nm Si wafer (see FIG. 29(a)) with mean value of 8.29×109 ohm-cm and standard deviation 1.30×107 ohm-cm;
FIG. 31(a) shows transverse resistivity from a MoAlOx film prepared by 10% Mo cycles beginning with Mo and Al2O3; FIG. 31(b) shows longitudinal resistivity;
FIG. 32 shows longitudinal and transverse resistivities for ALD of MoAlOx coatings measured at a field of 106 V/m versus % Mo cycles with film thickness indicated at each data point;
FIG. 33 shows temperature dependence of resistance of MoAlOx films of various % Mo;
FIG. 34(a) shows an SEM image of an ALD MoAlOx tunable resistance coating using 8% Mo ALD cycles in Al2O3 on an inside surface of a form borosilicate glass array and FIG. 34(b) shows a higher magnification image of a portion of FIG. 34(a).
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS The tuning of thin film electrical resistance can be done by mixing the conducting and insulating components in a precise controlled manner in composite materials. Schematic of such mixing and the resistivity (ρ) range of thin film composite materials is shown in FIG. 1. In the preferred embodiments the metal (Mo or W) and oxide (Al2O3) are combined by a ALD process under various deposition conditions. The materials can also be deposited by chemical vapor deposition type methods (CVD, MOCVD, OMCVD, PECVD) or PVD type methods such as molecular beam epitaxy (MBE), reactive sputtering, pulsed laser deposition (PLD), evaporation, etc.
The physical, electrical and chemical properties of nanocomposite thin films can be tuned by adjusting the relative proportions of the constituent materials. Amongst various thin film processes, atomic layer deposition (ALD) is a preferred technique for growing complex layers in a precisely controlled manner with many unique advantages. ALD is based on a binary sequence of self-limiting chemical reactions between precursor vapors and a solid surface. Because the two reactions in the binary sequence are performed separately, the gas phase precursors are never mixed; and this eliminates the possibility of gas phase reactions that can form particulate contaminants that might produce granular films. This strategy yields monolayer-level thickness and composition control. The self-limiting aspect of ALD leads to continuous pinhole-free films, excellent step coverage, and conformal deposition on very high aspect ratio structures. ALD processing is also extendible to very large substrates and batch processing of multiple substrates.
In one embodiment, thin films of W:Al2O3 nanocomposites were synthesized by combining tungsten (W) and aluminum oxide (Al2O3) using ALD processes. The ALD processes for both W and Al2O3 are known (also for Mo:Al2O3 to be discussed hereinafter. In addition, the ALD of W:Al2O3 nanolaminates comprised of alternating, distinct layers of these two materials has been explored, and these nanolaminates have been utilized as thermal barrier coatings and X-ray reflection coatings. In contrast to this previous work on nanolaminates, the instant invention in general focuses on synthesizing, characterizing, and testing nanocomposites where the W or Mo and the Al2O3 components do not exist in distinct layers but are more intimately mixed such that the domains of these materials do not exhibit bulk-like properties or structures (examples will be provided hereinafter). This offers the opportunity to develop very different materials with unique properties that are unlike any other constituents known in the art.
One exemplary use of these methods is to develop tunable resistive coatings using ALD for application in microchannel plate (MCP) electron multipliers. In this application, the ALD resistive layer serves to generate a uniform electrostatic potential along the MCP pores. The W:Al2O3 system was selected as a preferred article for a number of reasons. ALD W has a very low electrical resistivity of ρ=˜10−5 Ω-cm while ALD Al2O3 is an excellent insulator with a resistivity of ρ=˜1014 Ω-cm, and this contrast offers the potential for an extremely wide range of tunable resistance. In addition, ALD Al2O3 has a high breakdown electric field and this attribute is beneficial in high voltage devices such as MCPs. Both the W and Al2O3 ALD processes can be performed under similar conditions, and this simplifies the process of synthesizing composite layers. In addition to the wide variance in electrical properties, W and Al2O3 have very different physical and chemical properties. As a result, by adjusting the proportion of W in the Al2O3 matrix, we expect that the optical, mechanical, and physical properties of the W:Al2O3 composite layers can be broadly tuned.
Al2O3 ALD can be accomplished using alternating exposures to trimethyl aluminum (TMA) and H2O according to the following binary reaction sequence:
Al—OH*+Al(CH3)3→Al—O—Al(CH3)2*+CH4 (a)
Al—CH3*+H2O→Al—OH*+CH4 (b)
where the asterisks denote the surface species. Both (a) and (b) reactions are self-limiting and terminate after the consumption of all the reactive surface species. During reaction (a), TMA reacts with surface hydroxyl species, AlOH*, and deposits surface AlCH3* species liberating methane. In reaction (b), H2O reacts with the surface methyl species to form new Al—O bonds and rehydroxylate the surface while again liberating methane. Repeating the (a)-(b) reactions results in the linear ALD of Al2O3 films at a rate of ˜1.3 Å/cycle.
The ALD of single-element films requires a different surface chemistry than the surface chemistry employed for binary compounds like Al2O3. For W ALD, one of the reactants is a sacrificial species that serves as a place holder in the AB-AB-AB . . . binary reaction sequence. This sacrificial species is removed during the subsequent surface reaction. The W ALD film growth using Si2H6 and WF6 is accomplished by two self limiting surface reactions described in details and below as a quick reference:
WF3*+Si2H6→W—SiH*+SiHF3+2H2 a)
WSiH*+WF6→W—WF3+SiHF3 b)
where the asterisks denote the surface species.
During reaction (c), WF6 reacts with the sacrificial silicon surface species, WSiHyFz*, and deposits WFx species. In reaction (d), Si2H6 strips fluorine from the tungsten surface species, WFx*, and reforms the sacrificial silicon surface species. The reaction stoichiometry is kept undefined because the exact identity of the surface species is not known. Repeating these surface reactions (c) and (d), the W ALD growth occurs very linearly.
However initial ALD nucleation of Al2O3 layer on W surface or W layer on Al2O3 surface may not have same ALD growth behavior as pure Al2O3 on Al2O3 or W on W layer. The self limiting surface reaction discussed above changes as compared to pure Al2O3 and W due to mixing of the four precursors TMA, H2O, WF6, and Si2H6 during the growth of W:Al2O3 composite layers. These four precursors can be introduced on the substrate by a choice of precursor sequence in one or multiple ALD cycles. These precursors introduction can result in the presence of various functional groups such as —OH, —WF3, —SiHx, —Al(CH3), —AlFx on the already partially or fully deposited layer and can result in a complex ALD chemistry. In addition, during the growth of W:Al2O3 composite materials every ALD cycle, whether it is Al2O3 or W, the underneath layer may not have an appropriate surface species to facilitate the ALD growth. Further the reaction byproduct such as AlFx, HFx and CHFx may be involved in the ALD growth and may incorporate adverse effect on ALD growth. In addition, the nucleation delay for Al2O3 and W ALD film growth may affect the resulting film growth and the roughness. The roughness (surface area) of each individual layer can affect the nucleation of the W or Al2O3 process. Understanding nucleation is important for selected applications, especially when the ALD method is used to grow precisely controlled composite materials with desirable electrical properties. Various aspects of the ALD synthesis of W:Al2O3 composite layers were examined with variation of composition of the W:Al2O3 layers. The physical and electrical properties of W:Al2O3 composite layers are discussed hereinafter.
In a preferred method the W:Al2O3 composite layer depositions were carried out in a hot wall viscous flow reactor ALD reactor. The W:Al2O3 composite films were deposited on n-type Si(100), fused quartz, glass substrates and high aspect ratio (60:1) borosilicate glass micro-capillary plates. Prior to ALD process all the substrates degreasing was performed using a 10 min dip ultrasonic acetone cleaning. For Al2O3 growth, Al(CH3)3 [TMA] was obtained from Sigma-Aldrich with a 97% purity and deionized (DI) H2O vapor was used as a precursors. For W ALD, tungsten hexafluoride (WF6, Sigma Aldrich, 99.9%) and disilane (Si2H6, Sigma-Aldrich, electronic grade 99.995%) were used as precursors. All precursors were maintained at room temperature at ˜20° C. The background N2 flow was set to 300 sccm which gives a base pressure of 1.0 Torr in the ALD reaction chamber and was measured by a heated MKS Baratron 629B model. The precursors TMA and H2O were alternately pulsed in the continuously flowing N2 carrier flow using high speed computer controlled pneumatics valves in a desire ALD sequence. During TMA and H2O dosing, pressure transient increases of 0.2 Torr for TMA and 0.3 Torr for H2O when the reactants were introduced into N2 carrier flow nitrogen carrier flow. Similarly, WF6 and Si2H6 precursors were alternately injected into N2 carrier flow. During Si2H6 and WF6 dosing, pressure transient increases of 0.25 Torr for Si2H6 and 50 mTorr for WF6. The results utilized the optimized process condition and precursor's dose timing sequence for pure W and Al2O3 ALD processes. The main experimental conditions for ALD are summarized in Table 1.
TABLE 1
ALD experimental parameters and physical properties of various layers
Parameter Al2O3 W W:Al2O3
Deposition temperature (° C.) 150-300 100-300 200-300
ALD Precursors TMA, H2O Si2H6, WF6 TMA, H2O, Si2H6, WF6
Growth rate (Å/cycle) 1.1-1.3 4.5-5.5 1.2-1.5
Substrates Si(100), fused Si(100), fused Si(100), fused quartz, glass
quartz quartz micro-capillary plate
XRD analysis Amorphous Nano-crystalline Amorphous
ALD cycle timing (s) 1-5-1-5 1-5-1-5 W(1-5-1-5):Al2O3(1-5-1-5)
Resistivity (Ω-cm) ~1014 ~10−4 10−4-1014
An in-situ quartz crystal microbalance (QCM) study was performed for various ALD processes Al2O3, W and W:Al2O3 with a different mixture of ALD cycle ratios. QCM mass gains were recorded for each case. Front sided polished QCM sensors were obtained from Colorado Crystal Corporation. The QCM housing is located inside the uniformly heated reaction zone of the ALD flow tube reaction chamber. The Maxtek BSH-150 sensor housing was modified to provide a slow nitrogen purge of 10-20 sccm over the back of the quartz crystal sensor. This nitrogen purge prevents reactant gases from entering the QCM housing and depositing material on the back surface of the QCM sensor. By preventing this deposition, the QCM yields absolute mass measurements. In addition, the nitrogen purge allows QCM measurements during the ALD of conducting materials such as W. These conducting materials would otherwise electrically short the QCM sensor and prevent oscillation. Mass uptake data was recorded at each ALD cycle in cases of Al2O3, W and W:Al2O3 composites with various ALD cycles ratio.
The thicknesses of W:Al2O3 layers were determined using spectroscopic ellipsometry measurements on the Si monitor coupons. Annealing of W:Al2O3 composite layers were performed at 400° C. in 500 sccm flowing Ar condition for 4 hrs at pressure of 1 Torr. The film thicknesses were measured using ex-situ ellipsometry and supported by cross-section scanning electron microscopy (SEM) analysis and transmission electron microscopy (TEM). The microstructure and conformality of W:Al2O3 layer coatings on Si substrates and high aspect ratio glass micro-capillary plates were examined by cross-sectional scanning electron microscopy (SEM) model Hitachi 4700. The electrical I-V characteristics and thermal coefficient of resistance (TCR) of W:Al2O3 layers were measured using a Keithley Model 6487 pA/V source. Electrical measurements were done using either micro probes or Hg-probe contact method. The resistance stability test was performed for several days under constant applied potential in vacuum.
Prior to ALD growth W:Al2O3 composite layers, pure Al2O3 and W layers growth was studied by ALD. Thickness series samples were prepared and characterized. The in-situ quartz crystal microbalance (QCM) was performed during the pure Al2O3 and W ALD films growth.
QCM measurements were performed during Al2O3 ALD at 200° C. using TMA and H2O with the timing sequence: (1-10-1-10). These conditions were chosen from the ALD growth saturation studies. Under these conditions FIGS. 2(a)-2(d) explain data acquired from Al2O3 QCM study and thicknesses of actual Al2O3 samples grown Si(100) with ALD growth experimental conditions described in Table 1. FIG. 2(a) shows a mass add-on a QCM at various precursor steps in one Al2O3 ALD cycle. Calculated mass add-on the QCM for the several Al2O3 ALD cycles are plotted in FIG. 2(b) which illustrates a very linear increase mass as increasing number of ALD cycles. A linear least-square fit to the Al2O3 QCM mass data indicates an average Al2O3 growth rate of 37 ng/cm2/cycle. The calculated mass/ALD cycle from FIG. 2(b) are shown in FIG. 2(c). This data confirms the every Al2O3 ALD cycles deposited under these set of conditions are deposited approximately at the same mass/ALD cycle. A thickness series of samples of Al2O3 were deposited on Si(100). The ellipsometry measured thicknesses of Al2O3 layers on Si(100) vs. number of ALD cycles are shown in FIG. 2(d). A linear least-square fit to Al2O3 thickness data yields to a growth rate of 1.21 Å/cycle. From the Al2O3 mass/ALD cycle and growth rate on monitor Si(100) data, the calculated Al2O3 density=3.1 g/cm3 which is lower than the bulk density of crystalline Al2O3=3.6 g/cm3. This density variation can correlate to the X-ray amorphous nature of the 600 Å Al2O3 layer shown in FIG. 4. Both the mass uptake and the growth rate data for Al2O3 per ALD cycle are in good agreement with prior work.
ALD of W performed on QCM at 200° C. with the repetition of a ALD precursors cycle Si2H6 (dose)-N2(purge)-WF6(Dose)-N2(purge) with precursor timings of (1-10-1-10)s. These conditions were chosen from the ALD growth saturation studies. W growth directly on Si(100) substrate has shown poor adhesion whereas Al2O3 passivated Si(100) substrate shows good adhesion with W layer. FIGS. 3(a)-3(d) describe the QCM and actual thickness data of W samples grown on 10 nm Al2O3 passivated Si(100) substrates. Mass up-take on QCM during various steps in one W ALD cycles is shown in FIG. 3(a). Calculated mass add-on the QCM for the 15 steady state W ALD cycles are plotted in FIG. 3(b). This data confirms a very linear increase mass growth with increasing number of ALD cycles. A linear least-square fit to the W QCM mass data indicates an average W mass rate of 930 ng/cm2/cycle. The calculated mass/ALD cycle is shown in FIG. 3(c) which confirms that W deposited under this set of experimental conditions are depositing nearly equivalent mass for every ALD cycles is same. Thickness vs. number of ALD cycles for W layer is shown in FIG. 3(d). A linear least-square fit to W thickness data yields to a growth rate of ˜5 Å/cycle. From the mass/ALD cycle and the W growth rate data on Si(100) data, the calculated W density=18.6 g/cm3 which is slightly lower than the bulk density of crystalline W=19.3 g/cm3. This variation density can linked to the nano-crystalline nature of ALD grown W layer analyzed by XRD shown in FIG. 4. XRD pattern of the ˜500 Å W shows three broad little peaks appearance due to nano-size crystallinity in the range of 2θ=35°-45° and are associated with the presence of α-W and β-W phases in the ALD grown W layer. The both mass/ALD cycle and growth rate/ALD cycles data for W are in-line with prior work.
Prior to W:Al2O3 composite layers growth, the QCM data were collected for the steady state ALD growth of Al2O3—W—Al2O3 with the precursor sequence of n(TMA-H2O)-m(Si2H6—WF6)-n(TMA-H2O) where n and m are the desire number of cycles and shown in FIGS. 5(a)-5(d). Repeatability of these data sets was collected for few times and confirmed repeatable behavior. FIG. 5(a) represents the mass deposited during ALD Al2O3—W—Al2O3 at various precursor steps for precursor n(TMA-H2O)-m(Si2H6—WF6)-n(TMA-H2O). As expected a rapid increased in mass add-on was noticed during W ALD cycle. This is related to the high growth rate and for high density of the W. The calculated mass add-on for Al2O3 and W calculated for every ALD cycles in Al2O3—W—Al2O3 system and is shown in FIG. 5(b).
In FIG. 5(b) the initial Al2O3 ALD cycles stopped after reaching a steady state mass add-on of 37 ng/cm2/ALD cycle. However, a noticeable change was observed on the mass uptake at the very first W cycle on Al2O3 surface. This mass add-on is ˜130 ng/cm2 which is ˜7 times lower compared to the W mass uptakes ˜930 ng/cm2/ALD cycle under steady-state ALD growth condition which was discussed in the previous section. There was no significant mass add-on for the next 3 W cycles. This can associate with the change in functional group and the reactive sites for the very first W ALD cycle and subsequent W ALD cycles. After this a rapid mass add-on was noticed up to next 8th W ALD cycles and mass add-on reaches ˜800 ng/cm2/ALD cycle for few cycles followed by gradual mass drop to ˜700 ng/cm2/ALD cycle followed by gradual increases in the mass add-on up to ˜930 ng/cm2/ALD cycle, then mass add-on reached a steady-state mass/W ALD cycle. An initial 8 ALD W cycle and very rapid increase in the mass can correlate to island-type growth. These islands can form discontinuous W layers. After the critical island size, the collapse of island might have occurred and the formation of the continuous layer begins where the change in surface area as well as number of reactive sites may have a contribution to the mass-add on rate. The crucial point is the first W ALD cycle mass-uptake can be especially useful in a controlling lower % W in W:Al2O3 composite system. The first total 8-9 W ALD cycles express the amount of higher % W deposits into the W:Al2O3 composite system.
First cycles of Al2O3 mass uptake on W surface was noted to be high ˜68 ng/cm2, which is ˜1.8 times more than the steady state growth condition 37 nm/cm2/ALD cycle (FIGS. 2(a)-2(d)). In contrast, Al2O3 steady-state mass/ALD cycle on W surface was achieved within next 4-5 ALD Al2O3 cycles. These mass up-take numbers change drastically as relative percentage ratio of the W to Al2O3 ALD cycles was altered and the precursor's sequence. This will be discussed in the next section. This lower amount of mass add-on for initial W ALD cycles was associated with nucleation delay which is dependent on the available functional groups to complete the reaction. This nucleation delay is crucial for the “rule-of-mixture” for the composite materials growth by ALD. The precise mass-up take of W and Al2O3 not only defines the growth and physical properties but also strongly affected the electrical properties of the W:Al2O3 composite layers.
Alteration of electrical properties of W:Al2O3 composite layers are preferably accomplished by adjusting relative ratio of the W to Al2O3 ALD cycles. The QCM study was performed at 200° C. for various compositions of the W:Al2O3 (i.e. x % Al2O3:y % W) ALD cycles using precursors sequence n[x(TMA-N2—H2O—N2)+y(Si2H6—N2—WF6—N2)] where x and y can vary between 0-100; and n is adjusted as per the desired ALD cycles which relate to the thickness. The precursors were dosed for 1 s followed by a 10 s N2 purge. FIG. 6 describes the QCM mass add-on data for various compositions of W:Al2O3 system. Three trends were observed: i) within each composition a linear increase of cyclic mass add-on as per increase in ALD cycles, ii) gradual increase of mass add-on up to 50%-50% W to Al2O3 ALD cycles which is associated with the disorder in ALD growth due to mixing of more or less W or Al2O3 and iii) >50% W ALD cycles conditions provide an exponential increase in mass uptake which can be associated with the increased number of W cycles that can lead to the rapid increase in the W mass portion in the composite layer as shown in FIG. 5(b). A calculation of the mass/ALD for each % W ALD cycle condition from FIG. 6 is expressed in FIG. 7(a). The exponential function fits well to this data set which confirms the mass/per ALD cycles increased exponentially with increasing % W ALD cycles. The actual W mass added per composition of W:Al2O3 is calculated by taking an average by adding actual W mass add-on from 30 ALD cycles for each of the W:Al2O3 compositions and dividing by the number of only W ALD cycles and normalized with 100% W mass/ALD cycles (930 ng/cm2/ALD cycle). It is noteworthy that this is the actual amount of W in the W:Al2O3, and it did not increase even up to 50% W ALD cycles (in Al2O3 dominated region). Whereas in ≧50% W ALD cycles conditions (in W dominated region), actual W mass rapidly increased mass behavior and also showed the exponential increasing trend for mass/per ALD cycle.
Thickness series samples of W:Al2O3 system with % W ALD cycles variations were deposited under similar conditions on Si(100) and thicknesses measured spectroscopic ellipsometry. FIG. 8(a) represents a growth rate vs. % of W ALD and shows an exponential increasing behavior. Data from both FIG. 7(a) and FIG. 8(a) support each other. Besides of this, the thickness series samples were deposited for the (30% W:70% Al2O3) ALD cycle composition by adjusting simply the number of ALD cycles and are plotted in FIG. 8(b). A linear least-square fit to thickness data yields a growth rate of 1.33 Å/cycle for W:Al2O3.
As deposited pure Al2O3, pure W as well as composite W:Al2O3 layers were uniform and smooth over the 300 mm Si substrate area. These layers were analyzed by X-ray diffraction analysis shown in FIG. 9(a). All the W:Al2O3 composite layers were shown to have an X-ray amorphous diffraction pattern. A further (30% W:70% Al2O3) ALD cycle condition sample was annealed at 400° C. in 300 sccm Ar at 1 Torr for 4 hours which also shows an X-ray amorphous behavior as illustrated in FIG. 9(b). For this growth condition the microstructure and thickness were also verified with cross-section TEM analysis for an as deposited sample capped intentionally with the 60 Å Al2O3 as shown in FIGS. 10(a)-10(d).
Cross-section TEM samples were analyzed at Evans Analytical Group (EAG). TEM ready samples were prepared using the in-situ FIB lift out technique on an FEI Strata Dual Beam FIB/SEM. The samples were capped with a protective layer of carbon prior to FIB milling. The samples were imaged with a FEI Tecnai TF-20 FEG/TEM operated at 200 kV in bright-field (BF) TEM mode, high-resolution (HR) TEM mode, high-angle annular dark-field (HAADF) STEM mode and nano-beam diffraction (NBD) mode.
It is clear from the FIG. 10(a) cross-section TEM image that the surface of the sample is very smooth; and TEM thickness data are in good agreement with spectroscopic ellipsometry data. The film substrate interface has a very clean amorphous layer due to native Si(100)+few ALD cycles of Al2O3. Cross-section TEM data also confirms the thickness of an amorphous Al2O3 capping layer. TEM nano-beam diffraction (NBD) data was captured from the film region and shown in FIG. 10(b). The NBD image did not illustrate either a clear diffuse pattern that one can expect from amorphous Al2O3 or a crystalline pattern associated with that of W. The TEM data and XRD data are inline. A possible explanation of such a type of TEM bright field diffraction pattern is that the W:Al2O3 layer shows X-ray amorphous nature which can contain a mixture of amorphous and nano-crystalline materials. FIG. 10(c) represents a high-angle annular dark-field (HAADF) STEM image which shows a very uniform distribution of nano-size particles (bright white) homogeneously embedded in a continuous amorphous matrix. A very high resolution cross-section TEM image shown in FIG. 10(d) confirms the presence of 1-2 nm size particles dark spots (W) embedded in an amorphous matrix (Al2O3). Magnification of this image clearly shows the nano-crystallites at the dark spot locations. The XPS analysis of this layer confirms the W is in metallic state and is discussed in the next section. Further, the thickness measured by spectroscopic ellipsometery [see FIG. 8(b)] and cross-section TEM [see FIG. 10(a)] results are in agreement.
The elemental composition across the layer of the (30% W:70% Al2O3) ALD cycle condition case for an as deposited sample capped intentionally with the 60 Å Al2O3 is shown in FIG. 11. Composite analysis of this layer was investigated by X-ray photo electron spectroscopy (XPS). XPS depth profiles were obtained by alternating an acquisition cycle with a sputter cycle during which material was removed from the sample using a 4 keV Ar+ source. In order to eliminate crater wall effects, the data were acquired from a smaller region in the center of the sputter area. In addition, Zalar rotation was used to help minimize sample roughening during the sputtering process. The sputter rate is calibrated using SiO2. Note that the sputter rate of this material is likely different than that of SiO2. XPS measurements performed on PHI Quantum 2000 with X-ray source monochromated Alkα 1486.6 eV. The acceptance angle was ±23° and takeoff angle 45°. The analysis area was 1400 μm×300 μm and the sputter rate 39.8 Å/min (SiO2 Equivalent). XPS data is quantified using relative sensitivity factors and a model that assumes a homogeneous layer. The analysis volume is the product of the analysis area (spot size or aperture size) and the depth of information.
FIG. 11 shows the sputter-XPS depth profile, and after removal of an Al2O3 cap or top layer, the overall elemental composition of the W:Al2O3 composite layer is very uniform across the layer thickness. This sputter data also confirms the thickness of the sample (˜570 Å) and is good agreements with thickness measured by spectroscopic ellipsometery (˜570 Å) and cross-section TEM (550 Å). The coating is composed primarily of Al, W, O and F with lower but significant levels of C. A thin oxidized Si layer and a thin W/WSix layer were present between the coating and Si substrate. Calculated average atomic concentration from a uniform region is in the inserted table in the FIG. 11. Note that the ratio of Al (26.6%) to O (40.1%) concentration is 2:3 which is constant with the formation of the Al2O3. Secondly, impurities such as F and C are also present uniformly across the film thickness. The percentage of F concentration in an as deposited film is as high as 16.5% and 5.6% for C. It is noteworthy that the W atomic percentage from XPS is about 10.8% which is also consistent across the film thickness. This W atomic percentage for (30% W:70% Al2O3) ALD cycle condition sample is in good agreement with the calculated actual % W mass for (11.3%) by QCM mass add-on study for (33% W:66% Al2O3) ALD cycle condition sample and should be slightly higher due to 3% W more ALD cycles.
XPS spectra of individual elements were taken with respect to their binding energies after removing the W:Al2O3 composite layer with a sputter rate of (˜30 Å/min) and are shown in FIGS. 12(a)-12(f) and summarized in Table 2. The XPS spectral regions of Al 2p (FIG. 12a), O 1s (FIG. 12b), W4f (FIG. 12c), F 1s (FIG. 12d), C 1s (FIG. 12e) and Si 2p (FIG. 12f) signals collected across the sample. Spectral features of Al 2p with BE=75.8 eV and O 1s with BE=532.34 eV are very comparable across the films which confirms a uniform arrangement of Al2O3 surroundings for other elements within the film thickness. Amazingly, W 4f spectral region in FIG. 12(c) shows characteristically strong presence of metallic W throughout layer thickness. These XPS data supports the nanostructures observed in HAADF STEM and HRTEM images (FIG. 11) where 1-2 nm metallic W particles embedded homogeneously in Al2O3 amorphous matrix. The W 4f features of metallic W and tungsten carbides appear at 31.4 eV and 33.4 eV, while those of oxidized W are characterized by peaks at 35.5 and 37.7 eV. The W4f peaks show a small shoulder in between 36-38 ev range and can be associated with evidence of formation of small tungsten oxide that might have formed as a native oxide on W nanoparticles. The XPS spectra of C1s features has a peak position at 283.8 eV in FIG. 12(e) which shows a broadening with weak signal. The source of C signal is from —(CH3)* group form the TMA ALD precursor. We believed that this C1s position at 283.8 eV is mostly due to carbide formation. In this case C may have bonded to the W which can contribute to the very small fraction of formation of WC in the W:Al2O3 composite layer.
TABLE 2
Summary of XPS binding energy regions for W:Al2O3
XPS reference peak XPS Peak
position at BE (ev) positions form Possible elemental
Element taken from sample at BE (ev) bonding
Al2p 73 75.8 Al in Al2O3
O1s 531 532.3 O in Al2O3
W4f 31.4 31.6 Metallic W or WC
F1s 685 687.1 AlF3•3H2O
C1s 284.5 283.8 Carbide, WC
Si2p 99.3 99.4 Si(100) substrate
Evidently, F1s features in FIG. 12(d) are a strong signal and associated with AlF3.3H2O (aluminum fluoride trihydrate). Formation of AlF3.3H2O percentage might have a relationship to the ALD precursors, especially TMA and WF6 and precursor cycles sequence. This particular sample was prepared using (30% W:70% Al2O3) ALD cycle ratio using the ALD precursor cycle sequence of n[x(TMA-H2O)-y(Si2H6—WF6)]. Note that for n[x(TMA-H2O)-y(Si2H6—WF6)] ALD cycle, the WF6 precursor dose will leave reactive species *WF3 on the deposition surface; and on top of this we introduced TMA which can produce AlF3 as a reaction by-product which has high melting point, boiling and low volatility. Furthermore this AlF3 subsequently is hydrated as a result of a follow-up H2O precursor dose. This could explain the formation of AlF3.3H2O in the film with respect to use of ALD precursors and their sequence. Formation of AlF3.3H2O could also cause different mass uptake when introduced TMA after a WF6 precursor dose. This mass-add on is about twice the mass in comparison to TMA-H2O as shown in FIG. 5. As thickness of the layer is increased, AlF3.3H2O mass add-on could change. Nevertheless the AlF3.3H2O concentration will depend on n, x, y and ALD precursor sequence n[x(TMA-H2O)-y(Si2H6—WF6)] for any given ALD condition as a result of the nature of the precursor, especially TMA and WF6 that is used for W:Al2O3 ALD growth. After annealing of this composite layer the F amount is reduced to 7.8% compared to an as grown sample F amount (16.5%) shown in FIGS. 13(a) and (b) which represent XPS scans of as grown samples and after 400° C. annealing in Ar for 4 hours at pressure of 1 Torr.
The conformality of the composite W:Al2O3 layer was studied by depositing this layer on the high aspect ratio (AR=60) micro-capillary borosilicate glass plate shown in FIGS. 14(a)-(d). Cross-section SEM analysis confirms the layer was deposited conformally on the wall of the glass pores. As deposited pure Al2O3 (600 A) and W (500 A) layers grown on Si(100) show an AFM RMS roughness of 5.1 Å and 9.5 Å, respectively. This can be compared to a 650 Å (30% W-70% Al2O3) ALD condition composite sample which shows an RMS roughness of 14.1 Å on Si(100) substrate and 15.9 Å inside of glass capillary pore which is less than 2% of actual thickness of composite W:Al2O3 layer.
The precursor sequence in ALD has an important role, and this will define the functional group or reactive side for a sequential precursor dose. FIGS. 15(a)-(d) display the data from the QCM study performed at 200° C. for (25% W:75% Al2O3) ALD cycle case. This is a simple repetition of an ALD super cycle which includes [3x(TMA-N2—H2O—N2)+1x(Si2H6—N2—WF6—N2)] cycles up to the desired thickness. All precursors were dosed for 1 s and 10 s purge after every precursor dose shown in FIG. 15(a). A linear mass add-on behavior was observed for an ALD super cycle and within the super cycle. Magnification or zoom-in data of one of such a super cycle is shown in FIG. 15(b) which shows the actually mass add-on from each ALD cycle within one super ALD cycle. The calculated mass add-on/ALD cycle from the FIG. 15(a) is plotted as per the ALD cycles shown in FIG. 15(c). This shows liner mass add-on/ALD cycle. Mass uptake for every single ALD cycles is plotted against the number of ALD cycles. FIG. 15(d) shows four distinct mass add-on regions, with the bottom three being masses add-on due to 3x(TMA-N2—H2O—N2) ALD cycles. Top mass add-on was due to 1x(Si2H6—N2—WF6—N2) ALD cycles.
This mass add-on for one super ALD cycle or within the super cycle is significantly lower than the steady state mass uptakes for W and Al2O3 shown in FIGS. 2(a)-2(d) and 3(a)-3(d). The sum of mass uptakes for 3x(TMA-N2—H2O—N2) ˜63 nm/cm2 give an average of ˜21 nm/cm2 that is ˜45% lower than the steady state mass uptakes (37 ng/cm2/ALD cycle) for Al2O3. From the total mass of 3x(TMA-N2—H2O—N2) ˜63 nm/cm2 this corresponds to the ˜2 Å of Al2O3 thickness that corresponds to formation of ˜¾th monolayer of AlOx based on the physical thickness of AlOx monolayer=˜2.7 Å. For this particular ALD cycle condition the average mass adds-on/per W ALD cycles after every 3xAl2O3 cycles is ˜130 nm/cm2 which is ˜7 times lower than the W mass uptake for a steady state ˜930 ng/cm2/ALD cycle and gives a growth rate for W ˜5 Å. This is an equivalent thickness of W for mass ˜130 ng/cm2=˜0.7 Å which=⅕th of monolayer thickness of W, based on the physical thickness of monolayer of W ˜4 Å with equivalent mass ˜745 ng/cm2. It is clear that the W layer will not be continuous and will not form even a monolayer up to 95% of W ALD cycles condition and is supported by mass update data from FIG. 7(a). Nevertheless, this ˜0.7 Å layer could deposit clusters of W which may be sub-nanoparticles in size. We believe these QCM mass add-on calculations data and the existence of nano size W particles that we observed in HRTEM micrographs support each other. On the other hand, the mass uptakes are very linear and reproducible for each ALD super cycle.
Similar to the FIG. 15(a) data, the add-on for the various ALD precursors sequence for 25% W:75% Al2O3 ALD cycles growth were recorded and summarized as all-in-one in FIG. 16. The mass/per ALD for super ALD cycle are extracted for all the conditions and are summarized in Table 3. The major mass add-on occurs during exposure of TMA and WF6 precursors. Si2H6 exposure on H2O or H2O exposure on Si2H6 has shown very little effect on the mass add-on.
TABLE 3
Mass uptakes on QCM during various ALD cycle precursor sequence
Average Mass uptake
Layer Precursors s quence Notation (ng/cm2/ALD cycle)
Al2O3 1x(TMA-H2O) 37
W 1x(Si2H6-WF6) 930
3xAl2O3 1xW Total
W:Al2O3 3x(TMA-H2O)- THSW 63 129 192
1x(Si2H6—WF6)
W:Al2O3 3x(TMA-H2O)- THWS 52 136 188
1x(WF6—Si2H6)
W:Al2O3 3x(H2O-TMA)- HTSW 62 87 149
1x(Si2H6—WF6)
W:Al2O3 3x(H2O-TMA)- HTWS 53 89 142
1x(WF6—Si2H6)
A similar mass uptake was noticed for the ALD precursor sequence THWS and THSW and can relate to W growth on —OH rich surface. AlF3.3H2O can cause exposure of TMA on W which actually adds the mass shown in FIGS. 17(a) and (b). In contrast to ALD with THWS and THSW precursor sequence ˜25% less mass uptake rate was recorded when ALD was performed with a HTWS and HTSW precursor sequence in ALD cycle. This lesser mass uptake can be linked to the W growth on —Al(CH3) rich surface and to after W exposure of H2O precursor which did not add any mass in both HTWS and HTSW as shown in FIGS. 17(c) and (d).
Current-Voltage (I-V) characteristics of the W:Al2O3 with various compositions were measured. FIG. 18 represents the measured resistivity from an I-V analysis vs. % of W ALD cycles in W:Al2O3 composites. A wide span of resistivity was observed and resistivity decreases rapidly when % W ALD cycles increased in W:Al2O3 composite growth. This type of controlled tunable resistive coating was applied as a resistive layer for a microchannel plate and can be used in photodetection applications, as well as in other applications previously cited herein.
Transverse (⊥) characteristics were determined by appropriate contacts where the electric field is perpendicular and longitudinal (∥) characteristics where the electric field is parallel to the W:Al2O3 composite layers I-V. In the longitudinal measurement the high surface area microchannel plate was used and a gold pattern comb structure contained 80000 sq and 2 μm spacing. For transverse measurements a TiN deposited Si substrate was used which makes bottom contact; and a top contact was made with a Hg drop set-up with dot size of about 812 μm.
In FIG. 19 the transverse and longitudinal resistivities were plotted against the electric field for W:Al2O3 layer deposited with the 30% W ALD cycle condition sample from FIG. 18. Both transverse and longitudinal resistivity show slight differences. Annealing this sample at 400° C. in 300 sccm argon at 1 Torr for 4 hours resulted in both transverse and longitudinal resistivity decreasing by nearly one order of magnitude. In both as grown and annealed condition measurements transverse resistivity is slightly higher than the longitudinal resistivity. Resistive values vs. field shows more or less a flat response up to electric field ˜107V/m and follows the rapid decrease in resistivity as field increases which can correlate to a conduction mechanism in the layer.
The possibility of Fowler-Nordheim tunneling mechanism here is low because it normally requires very high electric field (E) for reasonable conduction which is about ˜1 GV/m. Instead of this conduction for the W:Al2O3 composite it is likely to occur through one of the two predominant conduction mechanisms for insulators, Frankel-Poole (FP) emission, which has the following form,
J ∝ E exp(−q(φb−(qE/πε)1/2/kBT)) (1)
or Space-Charge Limited (SCL) emission, which at high field is
J˜εμ(V2/L3) (2)
These two mechanisms have different IV behavior. At lower field the IV curve follows linear ohm's law (V=IR) where as the I-V curve of FP emission is characterized by a straight line at large E on a semi-log plot of J/E vs. E1/2. In contrast to this the IV curve of SCL emission has a second order dependence on E. In addition to this only FP emission will show the temperature dependence, whereas both SCL and FN tunneling will not.
FIGS. 20(a)-20(d)(2) represent electrical measurements for a W:Al2O3 layer grown with a 30% W ALD cycle condition, IV curve of a film in FIG. 20(a) for log(J/E) vs. log (E) 20(b), on a semi-log scale J/E vs. E1/2 in FIG. 20(c), and in 20(d) data fitting at low and high field with two different fits (at low field a linear least-square fit and exponential fit). Lower field data follows Ohm's law for J vs. E. At high electric fields the data forms a line which is consistent with the FP emission mechanism. To establish straight SCL emission characteristics, the data is plotted on a log-log scale FIG. 20(b) and J/E did not show 2nd order dependence on E. Also we noticed that the electrical behavior of the film was dependent on temperature (discussed later and shown in FIGS. 21(a) and (b)). As a result it is likely that the electron transport in W:Al2O3 composite layer is FP emission, and not SCL emission.
The overall picture of the electrical transport mechanism for the W:Al2O3 composite is likely due to W nano clusters/particles embedded in an amorphous Al2O3 matrix. It is unlikely that SCL emission or FN tunneling plays a significant role. The fundamental mechanism appears to be FP emission. The majority of the carriers are likely to be from the W nano-clusters, which have their Fermi level pinned at the Al2O3 defects and which are likely to reside on the W/Al2O3 interface. The ionization probability follows the Fermi-Dirac distribution, which is already incorporated into the FP emission equation. Nevertheless, these W nano clusters/particles embedded in a homogeneous amorphous Al2O3 matrix show a useful and distinctive conducting mechanism.
I-V behavior in the temperature range 30-130° C. were measured for (30% W:70% Al2O3) ALD cycle case sample. The resistance at different temperatures are shown in FIG. 21(a) and plotted per the known Steinhart-Hart equation. A linear fit to this data give a negative slop which is a temperature coefficient of resistance=−0.024 for The W:Al2O3.
The W:Al2O3 composite layer shows a negative temperature coefficient of resistance (PTCR) effect from 30° C. to 130° C. The PTCR effect is a very common phenomenon observed in semiconducting layers and defines the thermal runaway of the materials. FIG. 21(b) shows log(R) vs. 1/T plots for the W:Al2O3 composite layer films on comb structure wafer substrates. The IV data were collected from room temperature (300 K) to 403K. The films show resistivities of the order of 3.2 to 0.27 MΩ-cm over this temperature range. The straight-line nature of the Arrhenius plots indicates thermally activated conduction, as often found in semiconductors. From the slope of the curves the values of the activation energy (Ea) are obtained, which correspond to the minimum energy required to transfer electrons from the donor level to the conduction band. Values of Ea obtained are 110 meV. This thermal coefficient and activation energy data are useful in for thermistor and MEMS devices.
In another embodiment of the invention involving Mo:Al2O3 (or MoAlOx) the Mo:Al2O3 composite layer depositions were preferably carried out in a hot wall viscous flow reactor ALD reactor. The Mo:Al2O3 composite films were deposited on n-type Si(100), fused quartz, glass substrates and high aspect ratio (60:1) borosilicate glass micro-capillary plates. Prior to ALD processing all the substrates were degreased using a 10 min dip ultrasonic acetone cleaning. For Al2O3 growth, Al(CH3)3 [TMA] was obtained from Sigma-Aldrich with a 97% purity and deionized (DI) H2O vapor was used as a precursors. For W ALD, tungsten hexafluoride (MoF6, Alfa Aser, 99.9%) and disilane (Si2H6, Sigma-Aldrich, electronic grade 99.995%) were used as precursors. All precursors were maintained at room temperature at ˜20° C. The background N2 flow was set to 300 sccm which gives base pressure of 1.0 Torr in the ALD reaction chamber was measured by a heated MKS Baratron 629B model. The precursors TMA and H2O were alternately pulsed in the continuously flowing N2 carrier flow using high speed computer controlled pneumatics valves in a desire ALD sequence. During TMA and H2O dosing, pressure transient increases of 0.2 Torr for TMA and 0.3 Torr for H2O when the reactants were introduced into N2 carrier flow nitrogen carrier flow. Similarly, MoF6 and Si2H6 precursors were alternately injected into N2 carrier flow. During Si2H6 and MoF6 dosing, pressure transient increases of 0.25 Torr for Si2H6 and 50 mTorr for MoF6. The main experimental conditions for ALD are summarized in Table 4.
TABLE 4
Experimental conditions for Mo:Al2O3
No. Parameters Values
1 Precursor for Al2O3 TMA and H2O
2 Precursor for Mo Si2H6 and MoF6
3 Number of ALD cycles Varied depend on desire thickness
for Mo:Al2O3
4 Deposition temperature 100-400° C.
5 Relative ratio of Mo:Al2O3 Varied depend on the desire
resistivity
6 Monitor Substrates Quartz, Si(100), MCP, comb
structures, Mo, Au TiN
7 ALD cycles The purge and dose time adjusted
according to desire resistivity
The thicknesses of MoAlOx layers were determined using spectroscopic ellipsometry measurements on the Si monitor coupons. Annealing of MoAlOx composite layers were performed at 400° C. in 500 sccm flowing Ar condition for 4 hrs at pressure of 1 Torr. The film thicknesses were measured using ex-situ ellipsometry and supported by cross-section scanning electron microscopy (SEM) analysis and transmission electron microscopy (TEM). The microstructure and conformality of MoAlOx layer coatings on Si substrates and high aspect ratio glass micro-capillary plates were examined by cross-sectional scanning electron microscopy (SEM) model Hitachi 4700. The electrical I-V characteristics and thermal coefficient of resistance (TCR) of MoAlOx layers were measured using a Keithley Model 6487 pA/V source. Electrical measurements were done using either micro probes or Hg-probe contact method. The resistance stability test was performed for several days under constant applied potential in vacuum.
X-ray diffraction (XRD) analysis was performed to test deposited materials crystallinity or preferred any crystallographic phase. X-ray diffraction of as deposited MoAlOx layers deposited with 8% Mo ALD cycles in Al2O3 layer shows an amorphous structure (see FIG. 22).
To evaluated the microstructure, surface roughness and uniformity of the deposited MoAlOx materials. AFM analysis (shown in FIGS. 23(a) and (b)) of ALD MoAlOx tunable resistance coating prepared using 8% Mo cycles. with a thickness of 910 Angstroms on borosilicate substrate. The film surface is very smooth and shows only fine, nanoscale features. The RMS roughness is 0.634 nm for 91 nm film.
To evaluated the across the layer microstructure, uniformity and crystallinity of the deposited materials MoAlOx films were analyzed with TEM. Top down transmission electron microscopy (XTEM) image FIG. 24(a) was performed on an ALD MoAlOx tunable resistance coating prepared using 8% Mo cycles with a thickness of 760 Angstroms on Si substrate. This image shows a uniform layer film on Si substrate and the TEM bright field image FIG. 24(b) confirms an amorphous layer deposited under these conditions which is desirable for the many microelectronics applications.
The cross section transmission electron microscopy (XTEM) images of FIGS. 25(a)-(c) show an ALD MoAlOx unable resistance coatings prepared using 8% Mo cycles with a thickness of 760 Angstroms on Si substrate. These images illustrate the pinhole free film on Si substrate which contains nanoparticles of Mo embedded in amorphous Al2O3 matrix.
The MoAl2O3 processes were tested on the 300 mm Si wafer as well as 12″×12″ glass substrate shown in FIG. 26(a); and it shows excellent thickness uniformity across the wafer shown in FIG. 27(b) and details are given in Table 5. Thickness contour maps are shown in FIGS. 27(a)-27(f) for MoAl2O3 films prepared using 500 cycles with various % Mo and also shown are associated refractive index maps for each % Mo example.
TABLE 5
Summary of data shown in FIGS. 27(a) and 27(b) giving minimum,
maximum, average, and % standard deviation values for the thickness and
refractive index data in these plots.
Thickeness [A] Index of refraction
Mo %=> 5 8 12 5 8 12
Minimum 608 775 987 1.62 1.72 1.92
Maximum 619 794 1112 1.67 1.77 2.03
Average 613 785 1010 1.63 1.73 1.96
% STDV (1 σ) 0.52 0.57 1.82 0.48 0.84 1.35
Transverse electrical measurements of the MoAlOx layer were performed on the 300 mm Si wafer deposited first with Mo (served as bottom contact). The schematic of the test structures on 300 mm Si wafer is shown in FIG. 28. FIGS. 29-32 illustrate resulting data for these examples.
The MoAl2O3 layers were deposited with various thicknesses and compositions. The longitudinal and transverse resistivities were measured for several samples and plotted against the % of Mo ALD cycles in Al2O3 and shown in FIG. 32. Both transverse and longitudinal resistivities are more or less same within the experimental error limits.
I-V behavior in the temperature range 30-130° C. was measured for various Mo:Al2O3 layers The resistance at different temperatures are shown in FIG. 33 and plotted per [R=Roe−(βT(T−To))] A linear fit to this data give a negative slope which is a temperature coefficient of resistance. From these data we can extract the activation energy (Ea), which corresponds to the minimum energy required to transfer electrons from the donor level to the conduction band. Values of Ea obtained are given in last column of Table 6. This thermal coefficient and activation energy data are useful for thermistor and MEMS devices.
The temperature dependence of resistance was fit to:
R=Roe−(βBT(T−To))
Where R is resistance, T is temperature and Ro is initial resistance at initial temperature and βT is the thermal coefficient of resistance.
For reference:
- Commercial glass MCP: βT=−0.02
- Microchannel Silica MCP: βT=−0.036
- Aluminum zinc oxide coated MCP: βT=−0.06
TABLE 6
Summary of measurements of resistance of MoAlOx, films for different %
Mo. Column 6 lists the βT value, the thermal coefficient of resistance. βT decreases
with increasing % Mo. Slope and Ea are described in FIG. 33.
Sample # cycles Mo % Comments Thickness (A) βT Slope Ea (eV)
D20 130 7.7 10:120 = Mo:AlO 193 −0.03683 4509 0.39
D21 132 9.1 12:120 = Mo:AlO 201 −0.02800 4056 0.35
D22 135 11.1 15:120 = Mo:AlO 215 −0.00962 2987 0.26
D37 140 14.3 20:120 = Mo:AlO 230 −0.00112 1940 0.17
The growth of ALD MoAlOx layer on high aspect ratio 3D structures was also evaluated. ALD MoAlOx tunable resistance coatings were prepared using 8% Mo ALD cycles in Al2O3 on the inside surface of a porous borosilicate capillary glass array showing excellent thickness uniformity, conformality, and smoothness of the films
SEM images of ALD MoAlOx tunable resistance coating prepared using 8% Mo ALD cycles in Al2O3 on inside surface of porous borosilicate capillary glass array showing excellent thickness uniformity, conformality, and smoothness of the films.
The foregoing description of embodiments of the present invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the present invention to the precise form disclosed, and modifications and variations are possible in light of the above teachings or may be acquired from practice of the present invention. The embodiments were chosen and described in order to explain the principles of the present invention and its practical application to enable one skilled in the art to utilize the present invention in various embodiments, and with various modifications, as are suited to the particular use contemplated.
